Processor-Speed-In-kmh
Previously we have discussed the processor chip, of course using the unit hertz (Hz). The higher the Hz, does it affects the speed of the “process”? Maybe yes maybe no. Compared to us who drive, the faster km/h we can reach, does that also mean we get to our destination faster? I’m curious about the two units because they both have speed initials. Another interesting question is, what if we drive as fast as the processor? Is it possible? Let’s discuss this further.
Table of Contents
Epistemologies
A clock speed of 3.2 GHz means, the chip has the capability of 3.2 billion executions per second. 1 execution signifies 1 periodic cycle(cycle). As in the following figure, 6 Hz is equal to 6 cycles every 1 second.
As we know, Hz is used as a unit of frequency. While ‘v’ (velocity) or speed, has units of m/s (meters per second). These two units have a relation related to the wavelength λ — lambda. The third formulation is v = λ x f. Due to various limitations, I will give an estimate, not the actual speed. Through this approach, at least it can be estimated.
The market language that we often encounter is none other than the size of chip fabrication. For example, there are two brands of processor manufacturers, let’s call them A and B. Chip A works on the 10nm process, Chip B uses the 7nm process, and even some GPUs and TSMC from the 2 manufacturers have worked on the 5nm process. In general, nanometers represent the minimum width between the D-the drain channel and the S-the source channel of the transistor.
Evidence of technological advances is the ability of manufacturers to reduce transistors, the goal is so that the chip can contain more transistors. Call it the Qualcomm Snapdragon 865 Processor released in 2020 which has 10,300,000,000 transistors in 1 chip. That is, there are 1,287,500,000 transistors for each core, for a total of 8 cores. Cortex A77 speed 2.84 GHz, with 7nm fabrication.
Calculation
Let’s simulate calculating the wavelength in a way; fabrication size multiplied by the number of transistors in 1 core.
λ = 7 nm x 1,287,500,000 transistor
λ = 7 x 10^-9 . 1,287,500,000 = 9m
v = λ x f = 9m x 2,84GHz = 9 x 2,84 x 10^9 m/s
v = 25,6 x 10^9 m/s
His speed exceeds the speed of light! I can’t believe it myself either. However, this figure cannot be compared with the speed of light, because it is not apple to apple. Why can’t it be compared? Because the speed obtained is a combination of billions of transistors in a process, it is not a particle propagation speed.
Comparison
Next, we try to do an equivalent comparison. The first commercial chip, the Intel 4004 released in 1971 had 1 million transistors, a clock speed of 750 kHz, 10 µm fabrication, and 2,250 transistor units. We try to calculate the speed with the same method and formula.
λ = 10 µm x 2,250 transistor
λ = 10 x 10^-6 x 2,250 = 22.5 mm
v = λ x f = 22.5 mm x 750 kHz = 22.5 x 10^-3 x 750 x 10^3 m/s
v = 16,875 m/s = 60,750 km/h
Although in fact, the processor speed is easier when compared to units of Hz. From the comparison of the 2 chips, their speed is fairer, right?
About The Author
